The company makes a B2B process intelligence product and is building the applied AI layer behind it.
You will turn messy enterprise data into dependable, structured insights built on hosted foundation models, across retrieval, evaluation, orchestration and production quality.
Responsibilities
Ship and maintain AI services that convert raw company data into outputs the product can use.
Build retrieval, context assembly, structured extraction and multi-step flows on top of hosted language model APIs.
Write evaluation sets, benchmarks, regression suites and launch gates to track and raise model quality.
Watch live systems, look into incidents customers see, and make the stack faster, cheaper, more reliable and easier to recover.
Add guardrails, checks and fallback paths that cut errors and make model behavior more predictable and safer.
Let customer input and product results shape experiments and what gets built next.
Requirements
Typically 5 to 10 years building production software, applied AI or ML products.
Strong Python skills for building services, including asynchronous patterns.
Hands-on work putting hosted LLM APIs into customer-facing products, covering structured outputs, retrieval, tool use or agentic workflows.
You have owned AI features from deployment through monitoring, and handled production incidents and debugging.
Working knowledge of AI evaluation, benchmarks, CI/CD, guardrails and regression testing.
Systems thinking and clear communication, and the ability to iterate with customers as the product changes.
A bachelor's degree or higher in computer science, engineering or a related technical field.
Go or another typed language is helpful.
Experience with retrieval systems, embeddings, search, ranking or extracting information from complex documents is valuable.
Familiarity with PostgreSQL, Redis, React, TypeScript, GCP, Terraform, Docker and vector databases is useful.